- Email: [email protected]
Development of polyelectrolyte multilayer membranes to reduce the COD level of electrocoagulation treated high-strength wastewater
Author’s Accepted Manuscript Development of polyelectrolyte multilayer membranes to reduce the COD level of electrocoagulation treated High-strength wastewater Oishi Sanyal, Zhiguo Liu, Brooke M. Meharg, Wei Liao, Ilsoon Lee www.elsevier.com
PII: DOI: Reference:
S0376-7388(15)30175-7 http://dx.doi.org/10.1016/j.memsci.2015.09.011 MEMSCI13967
To appear in: Journal of Membrane Science Received date: 1 April 2015 Revised date: 11 August 2015 Accepted date: 6 September 2015 Cite this article as: Oishi Sanyal, Zhiguo Liu, Brooke M. Meharg, Wei Liao and Ilsoon Lee, Development of polyelectrolyte multilayer membranes to reduce the COD level of electrocoagulation treated High-strength wastewater, Journal of Membrane Science, http://dx.doi.org/10.1016/j.memsci.2015.09.011 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Development of polyelectrolyte multilayer membranes to reduce the COD level of electrocoagulation treated high-strength wastewater Oishi Sanyal1, Zhiguo Liu2, Brooke M. Meharg1, Wei Liao2 and Ilsoon Lee 1* 1
Department of Chemical Engineering and Materials Science, 2 Department of Biosystems and Agricultural Engineering Michigan State University, East Lansing, Michigan 48824, USA *To whom correspondence should be addressed: Email – [email protected]
Abstract This study focused on developing a membrane-based purification process, coupled with electrocoagulation (EC) as the pretreatment step, to reduce the COD level of an anaerobic digestion effluent. Commercial brackish water reverse osmosis (RO) membranes offer high COD removal but very low water fluxes. In an effort to address this issue, polyelectrolyte multilayer (PEM) membranes were fabricated by the surface modification of loose nanofiltration membranes using layer-by-layer assembly technique. The application of PEM membranes to treat wastewater effluents has not been explored in details. Two polyelectrolyte combinations were tried – the first one consisted of poly (diallyl dimethyl ammonium chloride) and poly (styrene sulfonate) while the second one consisted of poly (allylamine hydrochloride) and poly (acrylic acid). In comparison to commercial RO membranes, these membranes offered significantly higher fluxes, albeit with equivalent COD reduction. The effect of effluent properties like pH and composition, on the performance of these membranes has been discussed. The PEM films were characterized based on properties like thickness and surface charge, which directly affected the separation behavior of the membranes. For the first time, the combination of EC and PEM membranes has been tried out as a simple, energy-efficient two-step process for treating high-strength wastewater.
1
Keywords: High-strength wastewater; membrane filtration; COD reduction; layer-by-layer assembly; polyelectrolyte multilayer membranes
Introduction With rapid population increase and industrial growth, the fresh water resources are quickly being depleted. Improving the efficiency of water usage is critical to achieving global water sustainability. Recycle and reuse of municipal, industrial and agricultural wastewater are one of the key approaches to realize such efficiency improvement. Among different types of wastewater, high-strength wastewater with a significant amount of organic matter, solids, and nutrients requires more complicated processes to be treated and turned into clean water. In the present study, liquid effluent from an anaerobic digestion (AD) reactor was the target highstrength wastewater. Due to the chemical complexity of AD liquid effluent, Chemical Oxygen Demand (COD) was used as an indicator to evaluate the removal efficiency of different treatments [1, 2].
The combination of electrocoagulation (EC) and membrane technology has
attracted attention as an effective approach to treat high strength wastewater [3-6]. Compared to other physical and chemical treatments such as sedimentation, flocculation and activated carbon, EC technology presents a superior method to remove solids and pollutants from various wastewater streams. It has advantages of shorter retention time, better removal of smaller particles, no need of additional coagulation-inducing reagents, and smaller footprint [7]. A previous study demonstrated that a two-stage EC process showed high COD removal from AD effluent [8]. However, the EC treatment was still not enough to meet the drinking water standards. Thereby, in order to turn AD effluent into clean water, membrane filtration was
2
employed to purify the EC treated effluent. The overall wastewater treatment scheme is shown in Figure 1.
Figure 1. Flow diagram of the overall wastewater treatment process* *: Membrane filtration was the focus of this study.
Various membrane technologies such as microfiltration (MF), ultrafiltration (UF), and reverse osmosis (RO) have been implemented on EC treated wastewater to produce clean water [4, 9, 10].
Dense nanofiltration (NF) and RO membranes exhibit high solute removal
properties[11]. However, the permeability of the commercially available RO or dense NF membranes are extremely low, which leads to high operating costs. This study focused on the surface modification of NF 270 which is a “loose” NF membrane with high permeability. We 3
used the layer-by-layer assembly (LbL), which is an aqueous based thin film deposition technique, to modify the membrane surface. It was hypothesized that these membranes, on modification, would offer equal COD reduction as commercial RO membranes, but with higher solution permeability. The LbL technique was pioneered by Iler [12], and Decher et al. did a significant amount of research in this area to make LbL one of the most versatile thin film deposition techniques [13, 14]. It involves the deposition of polyelectrolytes which can interact via secondary molecular interactions such as ionic bonding, hydrogen bonding, hydrophobichydrophobic interactions etc. [15]. Polyelectrolyte multilayer (PEM) membranes, which are fabricated by the LbL assembly of polyelectrolytes on different commercial membrane surfaces, have been widely employed for ion rejection applications such as water softening or desalination [16-22]. Most of these studies were performed using a lab-based synthetic solution, which fails to emulate the complexities involved in real wastewater samples. Only a few studies applied the PEM-based modified membranes on real effluents such as industrial effluents and tanning effluents [23, 24]. In these studies, no detailed filtration performance and separation characteristics of the modified membranes have been reported. Besides, the combination of EC and PEM-based membrane filtration to treat high-strength agricultural wastewater also has not been studied. In order to elucidate the performance of PEM membranes two different sets of polyelectrolytes were selected and the resulting membranes were compared against a commercial RO membrane. The first set of polyelectrolytes comprised of the electrostatically crosslinked poly(diallyl dimethyl ammonium chloride) (PDAC) and poly(styrene sulfonate) (SPS), while the second set comprised of poly (allylamine hydrochloride) (PAH) and poly (acrylic acid) (PAA) that were covalently crosslinked by N-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC). COD
4
reduction and membrane permeability were used as the criteria to compare the performance of the PEM membranes with the commercial RO membranes.
2. Materials and Methods 2.1 Materials PDAC (MW 100,000-200,000, 20 wt% in water), SPS (MW 70,000), PAH (MW 900,000) and PAA (MW 15,000, 35 wt% in water) were purchased from Sigma Aldrich. 2-(NMorpholino) ethanesulfonic acid (MES) Buffer was also purchased from Sigma Aldrich. EDC was purchased from Fisher Scientific. Sodium chloride and potassium chloride crystals were procured from Avantor Performance Chemicals (Center Valley, PA). Three commercial membranes, NF 270, NF 90 and BW 30, from Dow Filmtec (Midland, MI), were used as the base membranes for this study. Among them, NF 270 and NF 90 are nanofiltration membranes, and BW 30 is a brackish water reverse osmosis membrane. All aqueous solutions were prepared using deionized (DI) water (>18.2 MΩ) supplied by a Barnstead Nanopure Diamond-UV purification unit equipped with a UV source and a final 0.2 µm filter. Unless specified all procedures were carried out at room temperature (25 °C). 2.2 The Dead End filtration set-up The HP 4750 stirred dead end cell (Sterlitech, Kent, WA) was connected to a nitrogen cylinder which acted as the pressure reservoir. The net volumetric capacity of the setup was around 300 ml. The unit was placed on a magnetic stirring plate. The filtrate was collected in a measuring cylinder. The flux was calculated by measuring the volume of water collected over a certain period of time and normalizing it with respect to the membrane area. The effective membrane area for this setup was 14.6 cm2.
5
2.3 Measurement of effluent properties The total solids (TS) measurement was done using the dry weight method. COD, total nitrogen (TN) and total phosphorus (TP) measurements were done using HACHTM standards methods [25]. Total carbon (TC) and inorganic carbon (IC) were measured using Shimadzu TOC-VCPN total organic carbon analyzer (Columbia, MD, US). Total organic carbon (TOC) was measured by subtracting the IC value from TC. 2.4 Experimental methods 2.4.1 Electrocoagulation (EC) treatment AD effluent was obtained from a 2500 m3 completely stirred tank reactor (CSTR) in the Anaerobic Digestion Research and Education Center (ADREC) at Michigan State University. A two-stage EC treatment was then carried out as described in a previous study [8]. Current level of 2A was applied to two sets of electrodes with anodic surface area of 62 cm2. The effective volume of the EC reactor was 0.5 L and retention time of 60 minutes was employed. The middle layer with relatively less turbidity was siphoned out as the feed for the second stage of EC, which applied same electricity conditions for another 40 minutes. Solutions were centrifuged at 236 g for 10 minutes before being used for the membrane filtration experiments. 2.4.2 Preparation of polyelectrolyte solutions and LbL assembly technique The bare NF 270 membranes were stored in DI water prior to the LbL deposition. The permeate sides of the membranes were covered with four alternate sheets of parafilm and aluminum foil prior to the LbL process. The required portion of the membrane was then cut out and used for the filtration experiments. This way, we were able to prevent the deposition of polyelectrolytes on both sides of the membrane to a large extent.
6
The LbL deposition process was carried out using a Carl Zeiss Slide Stainer which employs a robotic arm to move the sample to different solution baths. For all polyelectrolytes, the concentration was maintained at 10 mM. PDAC and SPS were prepared in 0.5M NaCl solutions and no pH adjustments were made to these solutions (the unadjusted pH values of the PDAC and SPS solutions were 6.6 and 6.0 respectively). For the PAH/PAA multilayer assembly, the pH of PAH and PAA were adjusted to 8.5 and 3.5, respectively, using 1M HCl/1M NaOH. The dipping time in each polyelectrolyte solution was set to 10 mins. After each polyelectrolyte dipping step, the substrates were rinsed with DI water for three
consecutive times (2 mins, 2
mins and 1 min). Following the deposition of one complete bilayer, the sample was sonicated for 2 mins in an ultrasonicator bath. The procedure was repeated till the desired number of bilayers was deposited. After LbL deposition, the membranes were soaked overnight in DI water. 2.4.3 Crosslinking of PAH and PAA with EDC The EDC solution was prepared in a 50 mM MES Buffer solution at a concentration of 50 mg/ml and pH of 5.5. Following the deposition of PAH/PAA multilayers on the membrane surface using the aforementioned LbL protocol, the modified membrane was dipped in the EDC solution using the slide stainer for 60 minutes with continuous agitation. The samples were then subject to three consecutive DI water rinsing steps, each of them being of 15 minutes duration, and then followed by 5 minutes of sonication. 2.4.4 Dead end filtration protocol The membranes were stored in DI water at least for 12 hours prior to usage. All membranes were first compacted by running pure DI water across them at a pressure of 10 bar. After two hours of compaction, the steady state value of the flux was noted down as the pure water flux of the membrane under consideration. Prior to using the wastewater, the solution was
7
pre-filtered using a 0.22 µm Millipore filter. This solution was then made to pass through the membranes at a transmembrane pressure (TMP) of 5 bar till the permeate volume reached 50 ml. The solution flux was calculated based on the time taken to filter 50 ml permeate across the given area of the membrane. The permeability and COD reduction values of various membranes which are presented in the later sections of this paper, were calculated based on the average of three replicates of each membrane. 2.5 Thin film characterization 2.5.1 Measurement of streaming potential The surface charge of the membranes were measured using the Brookhaven EKA Electro-kinetic analyzer (BI-EKA, Brookhaven Instrument Corp., Holtsville, NY) equipped with a rectangular cell and a clamp cell. A poly (methyl methacrylate) (PMMA) substrate was used as the reference. All streaming potential measurements were carried out using 1mM potassium chloride (KCl) as the electrolyte. The bypass lines and the cell were rinsed with KCl solution after being rinsed with DI water several times. At least three replicates were used for each of the membranes that were tested for streaming potential. In order to measure the membrane surface charge as a function of pH, the streaming potential tests were carried out with KCl solution at three different pH conditions (3, 7 and 9). The pH of these solutions was adjusted using 1M HCl/ 1M KOH. 2.5.2 Measurement of thickness The thickness of the PEM films was measured using a Dektak surface Profiler. Both types of PEM films were deposited on plain glass slides for the measurement. In order to reconfirm the values obtained from the profilometry tests, we also used the J.A Woollam M-44 Ellipsometer. The PEM films were deposited on gold coated glass slides (VWR International,
8
US) for the ellipsometry tests. Prior to the LbL deposition, the gold-coated glass slides as well as the plain glass slides were treated with O2 plasma for 20 minutes using a Harrick plasma cleaner (Harrick Scientific Corporation, Broading Ossining, NY) at 30 W RF power under 100 millitorr vacuum pressure. Immediately after the plasma treatment, the substrates were put in the slide stainer for the LbL process, as per the protocol described earlier (Section 2.4.1). The model for generic films was used for the ellipsometry measurements, assuming a refractive index of 1.5. The thickness was determined along several spots on the substrate and at least three replicates of each type of PEM film were used to get an average value.
. 3. Results and Discussions 3.1 EC treatment of diluted AD effluent The characteristics of AD liquid effluent and EC treated AD effluent (post vacuum filtration) have been listed in Table 1. After a two-stage EC treatment [8], a transparent solution was obtained and 97% of COD was removed, with the final COD level being 330 mg/L. Even though the EC treatment demonstrated excellent efficiency for solid and COD removal, the COD level in the effluent was still high, considering that the ultimate aim was to produce drinking water. Membrane-based filtration was therefore employed to reduce the COD of the EC treated AD effluent.
9
Table 1. Characteristics of the AD liquid effluent and filtered EC treated AD liquid effluent (post vacuum filtration) Parameters
AD liquid effluent*
EC-treated AD liquid effluent (post vacuum filtration) pH 7.5-8.0 9 TS (w/w %) 0.90 --1 COD (mg L ) 10017 330 TP (mg L-1) 340 1.18 TN (mg L-1) 1233 150 -1 TOC (mg L ) 2332 100 *: The AD liquid effluent for EC pretreatment was made by diluting the raw AD liquid digestate five times. 3.2 Performance of commercial membranes on the EC treated AD liquid effluent The three commercial membranes, NF 270, NF 90 and BW 30 were tested with the EC treated AD effluent and their performances were compared in terms of their permeability and COD reduction. The experimental data demonstrated that the NF 270 membrane had the highest pure water flux as well as solution flux among the three membranes, while BW 30 showed the highest COD removal (Figures 2 and 3). It has been reported that pH of the solution has significant influence on the separation behavior of the membranes [26]. The original pH of the EC treated AD effluent was around 9. The pH of the EC effluent was adjusted to 3 in order to evaluate the effect of pH on permeability and COD reduction. As shown in Figure 2, NF 270 membrane had higher COD reduction at acidic pH than alkaline pH, but the solution permeability was lower at acidic pH. (Figure 3a). In case of both NF 90 and BW 30 membranes, there was no significant difference in COD reduction under different pH conditions. The COD reduction of BW 30 membrane was higher than that of the NF 90 membrane. Both NF 90 and BW 30 had very similar solution permeability which was much lower than the solution permeability of NF 270 membrane. (Figure 3a). 10
The effect of pH on membranes, especially on NF 270 membrane, can be interpreted based on the membrane surface properties as well as the characteristics of the EC effluent. NF90 and BW30 are dense non-porous membranes as compared to NF 270 membrane, which is a “loose”, relatively more porous membrane. Even though all three membranes have positive charge at acidic conditions and negative charge at alkaline conditions, the responses of the NF 90 and BW 30 membranes to changes in effluent properties are different than that of NF 270 membrane. Natural organic matters (NOM) are present in almost all types of wastewater samples and the EC-treated AD effluent is no exception. TOC analysis confirmed the presence of a significant amount of organic matter in the EC effluent (100 mg/l). Among the various types of NOMs present in the solution, it has been reported that the humic compounds are largely responsible for the pH-dependent behavior of the effluent. [27-30]. As the humic compounds play critical roles as electron mediators in microbial electron transfer during anaerobic digestion [31], their concentration in the AD effluent was relatively high, and the EC treatment was not able to remove all of them. Consequently, a certain amount of them remained in the EC effluent. According to previous studies, humic compounds lose most of their charge under acidic conditions, resulting in their aggregation[29].
On the other hand, at alkaline pH these
compounds exhibit negative charge. For dense membranes like NF 90 and BW 30, the aggregation of humic compounds has little influence on their performance. Therefore no significant difference in performance (both COD removal and solution permeability) was observed for the NF 90 and BW 30 membranes, when tested under two different pH conditions. The BW 30 membranes showed the highest COD removal (92-95% ) with a COD level of
11
around 21 ppm in the final filtrate, but extremely low solution permeability of around 0.6 l/m2·hr·bar. For NF 270, the membrane surface properties as well as the EC effluent properties had significant influence on its performance. Under acidic conditions, the humic compounds were inclined to settle down on the membrane surface, resulting in lowering the permeability of NF 270 membrane at pH = 3 as compared to pH = 9 (Figure 3a). However, the deposition of humic compounds created an additional layer that prevented the passage of the remaining humic compounds and other fine particles in the effluent, which helped improve the COD reduction of the membrane (Figure 2). Under alkaline conditions, charge-based repulsion existed between the negatively charged humic compounds and the membrane surface. This repulsion was not enough to lower the COD level down to what was achieved under acidic conditions, but helped to maintain a higher solution permeability. Considering the fact, that a loose surface structure provides some flexibility for the LbL coating experiments, NF 270 was selected as the underlying substrate for fabricating the PEM membranes.
12
Figure 2. Comparison in COD reduction between the commercial membranes. (TMP = 5 bar)
(a)
13
(b) Figure 3. Permeability comparisons between the commercial membranes based on (a) solution tested under two different pH conditions and (b) pure water.
3.3 Performance of PEM-based membranes on EC-treated AD effluent 3.3.1 [PDAC (0.5 M NaCl)/SPS (0.5M NaCl)] 5.5 multilayer system Both PDAC and SPS are strong polyelectrolytes, which remain charged over a wide range of pH. 0.5 M NaCl was used as the supporting electrolyte which helps to increase the surface charge density [32]. The surface charge plays an important role in shaping the rejection behavior of these polyelectrolyte-based membranes. Under acidic conditions, as the humic compounds lose most of their charge, effect of the charge on the outermost layer is minimal. It was, however, important to understand how the surface charge affects the rejection under alkaline conditions, when the humic compounds are negatively charged. Two different systems were investigated to treat the EC effluent at pH 9: the first one comprising of 5 bilayers of
14
PDAC/SPS ending with SPS (negatively charged), and the second one consisting of 5.5 bilayers of PDAC/SPS ending with PDAC (positively charged). The experimental data demonstrated that the 5.5-bilayer system ending with PDAC had higher COD reduction (~ 84 %) than the system ending with SPS (~70 %) (Figure 4). Under alkaline conditions, the negatively charged humic compounds bind to the positively charged PDAC layer and form an additional barrier layer that helps enhance the rejection, while, the system ending with a SPS layer mainly uses charge-based repulsion to reject the negatively charged humic compounds. Therefore, under alkaline conditions, the barrier layer formation by humic compounds on a positively charged surface is more instrumental in reducing the COD level, than charge-based repulsion that occurs with a negatively charged outermost layer. Based on the performance results, the multilayer membrane ending with PDAC was chosen over the one ending with SPS, and used for the following experiments. The number of bilayers is also a tunable parameter for the LbL process. Following some preliminary experiments, 5.5 bilayers was found to be optimum in order to achieve both high flux as well as high rejection.
15
Figure 4. COD reduction of two PDAC/SPS based modified membranes. (TMP = 5 bar) (-) Bare NF270 or Multilayer assembly on NF270 ending with negatively charged SPS outer layer. (+) Multilayer assembly on NF270 ending with positively charged PDAC outer layer.
The performance of the PDAC/SPS–based membranes was further compared with the commercial membranes. These modified membranes showed around 84% COD reduction at pH = 9 and 90% at pH = 3 with corresponding COD values of 33 and 51.5 ppm in the filtered solution (Figure 5), which were significantly better than the original NF270. Permeability experiments demonstrated that the pure water permeability of the PDAC/SPS coated NF 270 membrane was much higher than NF 90 and BW 30 membrane (Figure 6b). The modified membrane also showed approximately 3 times higher solution permeability than BW 30/NF 90 at pH 3 and 5 times higher at pH 9 (Figure 6a). Similar to the underlying NF 270 membrane, the PDAC/SPS based membranes also had lower solution permeability at acidic pH as compared to alkaline pH (Figure 6a).
As mentioned before, the formation of a barrier layer of humic
compounds on the membrane surface was a key factor influencing the flux. However, the attachment mechanisms of the barrier layer under acidic and alkaline conditions are different. Under acidic conditions, the layer is formed due to the aggregation and settling of the humic compounds on the membrane surface. Under alkaline conditions, the negatively charged humic compounds adhere to the positively charged PDAC surface by electrostatic attraction. Given that the flux is lower under acidic condition, we presume that the barrier layer formed at pH 3 was thicker and denser than what was formed at pH 9. Owing to the thicker barrier layer, the rejection of the membrane was higher under acidic condition. The deposition of PDAC/SPS multilayers led to a considerable enhancement in the COD reduction compared to the bare NF 270 membrane. The percentage reduction was almost equal to that of BW 30 membrane under 16
acidic conditions but lower than BW 30 under alkaline conditions. Covalently cross-linked multilayer structure was therefore studied in order to further improve the membrane performance.
3.3.2 [PAH (pH 8.5)/PAA (pH 3.5)] 5.5 multilayer system Covalent cross-linking enables to create a tightly woven network, which is capable of blocking unwanted components of wastewater much more efficiently than ionic cross-linking. Therefore, a covalently cross-linked multilayer structure was coated on the surface of the NF 270 membrane with PAH and PAA as the surface modifiers. Both PAH and PAA are weak polyelectrolytes that exhibit pH-tunable behavior. When PAH is at pH 8.5 and PAA at pH 3.5, they form thick loopy films with inter-penetrated layers [33]. The amine groups of PAH and the carboxylic groups of PAA can be reacted to form covalent linkages between them. This can be simply done by heating the multilayers up to a high temperature (>180 °C) [34]. However, as the underlying NF 270 membrane cannot tolerate such high temperature, EDC induced chemical cross-linking was applied. [35]. A high EDC concentration of 50 mg/ml was used to ensure high cross-linking density [35]. In order to enable comparison with the PDAC/SPS –based membrane, the same number (5.5) of PAH/PAA bilayers were deposited. The performance of the PAH/PAA-based modified membrane has been presented in Figures 5 and 6. The COD reduction was between 91-95% under both acidic and alkaline conditions (Figure 5), and the final COD level in the permeate stream was almost equal to that of BW 30 membrane This could be attributed to a tight dense polymeric network that was formed by crosslinking PAH (pH 8.5) and PAA (pH 3.5), which was capable of completely
17
blocking the passage of unwanted components. The pure water permeability of the PAH/PAA membrane was the lowest among all the membranes (Figure 6b), while the solution permeability under both pH conditions was higher than BW 30 and NF 90 membranes (Figure 6a). The thick layers of PAH/PAA coating contributed to the low pure water permeability of the modified membrane. In spite of having a lower pure water permeability, the PAH/PAA coated NF 270 membrane showed 1.7 times higher solution permeability than the BW 30 membranes. We believe this was due to lower fouling propensity of the PAH/PAA coated membrane in comparison to the commercial RO membranes. This has been discussed in more details in a later section. In addition, because of the tight dense polymeric network and the thickness of the coated layers, the deposition of an additional layer of humic substances under acidic pH did not contribute much to the rejection. Therefore, there were no significant differences in COD reduction and solution permeability under acidic and alkaline conditions.
The PAH/PAA
multilayer membrane had the same COD reduction as the PDAC/SPS system at acidic pH, but higher reduction at alkaline pH The flux of this system was however, lower than the PDAC/SPS system. As the performance of the PAH/PAA coated NF 270 membrane was independent of the pH of the wastewater, it eliminates the need to acidify the solution prior to using it. This, in turn, helps to prevent the severe organic fouling that takes place under acidic conditions.
18
Figure 5. Comparison in COD reduction between the modified membranes and the commercial membranes at different pH conditions. (TMP = 5 bar)
19
(a)
(b) Figure 6. Permeability comparisons between the modified membranes and the commercial membranes based on (a) the solution at two different pH conditions and (b) pure water 20
The ratio of solution flux over pure water flux could be used as a criterion to evaluate the anti-fouling property of different membranes. A higher ratio indicates better fouling resistance offered by the membrane surface[36]. The PEM membranes (PDAC/SPS and PAH/PAA) showed better anti-fouling properties than their commercial counterparts (Table 2).The creation of anti-fouling surfaces by the deposition of polyelectrolytes has been reported earlier as well [37-40]. Among all the membranes, the PAH/PAA-based membranes had the best anti-fouling performance under both pH conditions.
Table 2. The ratio of the average solution flux to the average initial flux for commercial membranes and the PEM membranes* Membranes
Js (50)/Jw (0) at pH = 3
Js(50)/Jw(0) at pH = 9
NF 270
0.201
0.382
NF 270 modified with (PDAC/SPS)5.5 NF 270 modified with (PAH/PAA)5.5 NF 90
0.315
0.500
0.678
0.829
0.125
0.157
BW 30
0.209
0.193
Js(50) : The solution flux of the membrane measured by calculating the time taken to filter 50 ml of the permeate solution Jw(0) : The pure water flux or the initial flux of the membrane *: Both the fluxes were normalized with respect to the pressure applied during the measurements
21
With the combination of EC and membrane filtration, the color of the AD effluent was completely removed (Figure 7). This is important from the point of view of producing potable water. The EC process greatly removed the majority of the COD and TS in the AD effluent. The following membrane filtration step, irrespective of the membrane used, further decolored and cleaned the EC treated AD effluent. A relatively less energy-demanding method for treating high-strength wastewater was therefore established.
Figure 7. Color change of the solution following different treatment stages Left to right: AD effluent, EC effluent and membrane filtrate.
3.4 Effect of the coating properties on the membrane performance 3.4.1 Thickness of the PEM coatings The thickness of the coatings plays an important role in determining the permeability of the modified membrane. The strong polyelectrolytes PDAC and SPS are fully charged irrespective of pH conditions. The strong ionic linkages between the positively charged PDAC backbones and negatively charged SPS lead to the formation of very thin flat coatings. The 22
addition of NaCl increases the surface charge density and also increases the thickness of the films [20]. PAH and PAA exhibit pH dependent ionization behavior. PAH is around 50% ionized at a pH of 8.5, and PAA is only about 10% ionized at pH of 3.5 [41]. Since they are only partially charged, they arrange themselves in a coiled conformation,
with a high level of
interlayer diffusion [33], which leads to the formation of thick loopy films.. Crosslinking reduced the swelling tendency of the polyelectrolytes and in turn also led to a decrease in the film thickness.[42, 43]. The thickness values of the two types of polyelectrolyte systems have been shown in Table 3. It is apparent that the thickness of the coatings directly affects the permeability. The thicker PAH/PAA coatings had lower initial permeability than the PDAC/SPSbased coatings.
Table 3. The thickness values of the two types of polyelectrolyte coatings on NF 270 membrane Type of coating
Thickness (nm)
[PDAC (0.5M NaCl) / SPS(0.5M NaCl)] 5.5
28.70 ± 4.66
[PAH (pH 8.5) / PAA (pH 3.5)]5.5
134.15 ± 9.87
3.4.2 Surface charge of the PEM coatings Surface streaming potential of membranes plays an important role in determining the membrane rejection properties. The streaming potentials of all three commercial membranes as well as two modified membranes under three different pH conditions (acidic, neutral and alkaline) are listed in Table 4.
23
Table 4. The streaming potential data for the commercial membranes and the modified membranes as a function of pH* Surface streaming potential (mV) @ pH 3
Surface streaming potential (mV) @ pH 7
Surface -streaming potential (mV) @ pH 9
NF 270
5.54 ± 0.13
- 4.17 ± 1.89
-21.35 ± 0.36
NF 270 modified with (PDAC/SPS)5.5 NF 270 modified with (PAH/PAA)5.5 NF 90
20.38 ± 1.66
29.24 ± 2.45
20.07 ± 5.86
20.50 ± 4.47
12.83 ± 2.30
10.26 ± 3.36
20.91 ± 0.99
- 15.74 ± 0.09
-15.33 ± 0.23
BW 30
12.70 ± 3.80
- 4.70 ± 1.45
-3.66 ± 0.45
Membrane
*: Data represent average of three replicates with standard deviation.
The PEM membranes remain positively charged throughout the entire pH range of 3-9, unlike the commercial membranes. For the PDAC/SPS system, the outermost PDAC layer remained almost fully ionized within a wide range of pH and therefore imparted positive charge to the membrane surface. This leads to a small variation in its performance under different pH conditions (Table 4). For the PAH/PAA system, the effect of pH is much more complicated since both polyelectrolytes have pH dependent ionization. When exposed to a pH of 3, the outermost PAH layer was fully ionized and therefore had the highest positive charge. With the decrease in percentage ionization, the surface streaming zeta potential also decreased as the pH was increased (Table 4). However, many other complicated phenomena are associated with this polyelectrolyte system, like the inter-diffusion of polyelectrolytes [44] and shifting of the pKa value of the polyelectrolytes [41] when incorporated within the multilayers.. Besides, the effect of cross-linking on the ionization of the weak polyelectrolytes not been studied till date. These
24
factors can individually or collectively influence the surface streaming potentials under varying pH conditions.
4. Conclusion With this study it can be concluded that integration of electrocoagulation and PEM-based membrane separation provides an efficient way to reclaim a high-strength wastewater. PEMbased membranes demonstrated excellent performance in treating an actual wastewater effluent. A comprehensive study of several key concepts involving the membrane surface properties as well as the effluent properties was performed to deal with the complexities involved in wastewater treatment. The pH of the wastewater solution played a pivotal role in determining the performance of the membranes mainly due to the NOMs present in the EC effluent. Both PDAC/SPS and PAH/PAA based membranes had higher solution permeability than the commercial reverse osmosis membrane (BW 30). The PDAC/SPS based membrane showed equivalent COD reduction as BW 30 under acidic conditions, while the PAH/PAA based membranes showed equal reduction as BW 30, irrespective of the pH conditions used. Both the polyelectrolyte-based membranes had much lower fouling propensity than the commercial membranes, which expands the scope of applicability of PEM membranes beyond some simple ion-rejection studies. List of Figure Captions 1. Flow diagram of the overall wastewater treatment process* *: Membrane filtration was the focus of this study 2. Comparison in COD reduction between commercial membranes. (TMP = 5 bar) 3. Permeability comparisons between the commercial membranes based on (a) solution tested under two different pH conditions and (b) pure water
25
4. COD reduction of two PDAC/SPS based modified membranes (TMP = 5 bar) (-) Bare NF 270 or Multilayer assembly on NF 270 ending with negatively charged SPS outer layer (+)Multilayer assembly on NF 270 ending with positively charged PDAC outer layer 5. Comparison in COD reduction between the modified membranes and the commercial membranes at different pH conditions. (TMP = 5 bar) 6. Permeability comparisons between the modified membranes and the commercial membranes based on (a) the solution at two different pH conditions and (b) pure water. 7. Color change of the solution following different treatment stages Left to right: AD effluent, EC effluent and membrane filtrate
Acknowledgment The authors sincerely thank Prof. Volodymyr Tarabara from the Department of Civil and Environmental Engineering at MSU for allowing the usage of EKA Electrokinetic analyzer in his lab. The authors would also like to acknowledge the DOD Strategic Environmental Research and Development Program (DOD SERDP W912HQ-12-C-0020) for funding the entire research work.
References [1] H.A. Moreno-Casillas, D.L. Cocke, J.A.G. Gomes, P. Morkovsky, J.R. Parga, E. Peterson, Electrocoagulation mechanism for COD removal, Separation and Purification Technology, 56 (2007) 204-211. [2] K. Yetilmezsoy, F. Ilhan, Z. Sapci-Zengin, S. Sakar, M.T. Gonullu, Decolorization and COD reduction of UASB pretreated poultry manure wastewater by electrocoagulation process: A posttreatment study, Journal of hazardous materials, 162 (2009) 120-132. [3] M. Ben-Sasson, Y.M. Lin, A. Adin, Electrocoagulation-membrane filtration hybrid system for colloidal fouling mitigation of secondary-effluent, Separation and Purification Technology, 82 (2011) 63-70. [4] M. Ben-Sasson, Y. Zidon, R. Calvo, A. Adin, Enhanced removal of natural organic matter by hybrid process of electrocoagulation and dead-end microfiltration, Chemical Engineering Journal, 232 (2013) 338-345. [5] V. Mavrov, S. Stamenov, E. Todorova, H. Chmiel, T. Erwe, New hybrid electrocoagulation membrane process for removing selenium from industrial wastewater, Desalination, 201 (2006) 290-296. 26
[6] V. Vinduja, N. Balasubramanian, Electrocoagulation-integrated hybrid membrane processes for the treatment of tannery wastewater, Environmental Science and Pollution Research, 20 (2013) 7441-7449. [7] M.Y.A. Mollah, R. Schennach, J.R. Parga, D.L. Cocke, Electrocoagulation (EC)- science and applications, Journal of hazardous materials, 84 (2001) 29-41. [8] Z. Liu, D. Stromberg, X. Liu, W. Liao, Y. Liu, A new multiple-stage electrocoagulation process on anaerobic digestion effluent to simultaneously reclaim water and clean up biogas, Journal of hazardous materials, 285 (2015) 483-490. [9] X. Chen, H. Deng, Removal of humic acids from water by hybrid titanium-based electrocoagulation with ultrafiltration membrane processes, Desalination, 300 (2012) 51-57. [10] W. Den, C.-J. Wang, Removal of silica from brackish water by electrocoagulation pretreatment to prevent fouling of reverse osmosis membranes, Separation and Purification Technology, 59 (2008) 318-325. [11] C. Bellona, J.E. Drewes, P. Xu, G. Amy, Factors affecting the rejection of organic solutes during NF/RO treatment—a literature review, Water Research, 38 (2004) 2795-2809. [12] R.K. Iler, Multilayers of colloidal particles, Journal of Colloid and Interface Science, 21 (1966) 569-594. [13] G. Decher, Fuzzy nanoassemblies: Toward layered polymeric multicomposites, Science, 277 (1997) 1232-1237. [14] G. Decher, J.D. Hong, J. Schmitt, Buildup of Ultrathin Multilayer Films by a Self-Assembly Process .3. Consecutively Alternating Adsorption of Anionic and Cationic Polyelectrolytes on Charged Surfaces, Thin Solid Films, 210 (1992) 831-835. [15] I. Lee, Molecular Self-Assembly: Smart Design of Surface and Interface via Secondary Molecular Interactions, Langmuir, 29 (2013) 2476-2489. [16] S.U. Hong, O.Y. Lu, M.L. Bruening, Recovery of phosphate using multilayer polyelectrolyte nanofiltration membranes, Journal of membrane science, 327 (2009) 2-5. [17] S.U. Hong, R. Malaisamy, M.L. Bruening, Optimization of flux and selectivity in Cl-/SO42separations with multilayer polyelectrolyte membranes, Journal of membrane science, 283 (2006) 366-372. [18] S.U. Hong, R. Malaisamy, M.L. Bruening, Separation of fluoride from other monovalent anions using multilayer polyelectrolyte nanofiltration membranes, Langmuir, 23 (2007) 17161722. [19] L. Krasemann, B. Tieke, Selective ion transport across self-assembled alternating multilayers of cationic and anionic polyelectrolytes, Langmuir, 16 (2000) 287-290. [20] L. Ouyang, R. Malaisamy, M.L. Bruening, Multilayer polyelectrolyte films as nanofiltration membranes for separating monovalent and divalent cations, Journal of membrane science, 310 (2008) 76-84. [21] O. Sanyal, I. Lee, Recent progress in the application of Layer-by-layer assembly to the preparation of nanostructured ion-rejecting water purification membranes, Journal of Nanoscience and Nanotechnology, 14 (2014) 2178-2189. [22] O. Sanyal, A.N. Sommerfeld, I. Lee, Design of ultrathin nanostructured polyelectrolytebased membranes with high perchlorate rejection and high permeability, Separation and Purification Technology, 145 (2015) 113-119. [23] U.K. Aravind, B. George, M.S. Baburaj, S. Thomas, A.P. Thomas, C.T. Aravindakumar, Treatment of industrial effluents using polyelectrolyte membranes, Desalination, 252 (2010) 2732. 27
[24] M.S. Baburaj, C.T. Aravindakumar, S. Sreedhanya, A.P. Thomas, U.K. Aravind, Treatment of model textile effluents with PAA/CHI and PAA/PEI composite membranes, Desalination, 288 (2012) 72-79. [25] A.M. Jirka, M.J. Carter, Analytical Chemistry, 47 (1975) 1397. [26] G. Capar, L. Yilmaz, U. Yetis, Reclamation of acid dye bath wastewater: effect of pH on nanofiltration performance, Journal of membrane science, 281 (2006) 560-569. [27] S. Hong, M. Elimelech, Chemical and physical aspects of natural organic matter (NOM) fouling of nanofiltration membranes, Journal of membrane science, 132 (1997) 159-181. [28] C.Y. Tang, Y.-N. Kwon, J.O. Leckie, Fouling of reverse osmosis and nanofiltration membranes by humic acids-Effects of solution composition and hydrodynamic conditions, Journal of membrane science, 290 (2007) 86-94. [29] W. Yuan, A.L. Zydney, Effects of solution environment on humic acid fouling during microfiltration, Desalination, 122 (1999) 63-76. [30] W. Yuan, A.L. Zydney, Humic acid fouling during microfiltration, Journal of membrane science, 157 (1999) 1-12. [31] L. Shao, T. Wang, T. Li, F. Lu, P. He, Comparison of sludge digestion under aerobic and anaerobic conditions with a focus on the degradation of proteins at mesophilic temperature, Bioresource technology, 140 (2013) 131-137. [32] B.W. Stanton, J.J. Harris, M.D. Miller, M.L. Bruening, Ultrathin, multilayered polyelectrolyte films as nanofiltration membranes, Langmuir, 19 (2003) 7038-7042. [33] S.S. Shiratori, M.F. Rubner, pH-dependent thickness behavior of sequentially adsorbed layers of weak polyelectrolytes, Macromolecules, 33 (2000) 4213-4219. [34] J. Park, J. Park, S.H. Kim, J. Cho, J. Bang, Desalination membranes from pH-controlled and thermally-crosslinked layer-by-layer assembled multilayers, Journal of Materials Chemistry, 20 (2010) 2085-2091. [35] P. Schuetz, F. Caruso, Copper-Assisted Weak Polyelectrolyte Multilayer Formation on Microspheres and Subsequent Film Crosslinking, Advanced Functional Materials, 13 (2003) 929-937. [36] J. Pieracci, J.V. Crivello, G. Belfort, Photochemical modification of 10kDa polyethersulfone ultrafiltration membranes for reduction of biofouling, Journal of membrane science, 156 (1999) 223-240. [37] W. Shan, P. Bacchin, P. Aimar, M.L. Bruening, V.V. Tarabara, Polyelectrolyte multilayer films as backflushable nanofiltration membranes with tunable hydrophilicity and surface charge, Journal of membrane science, 349 (2010) 268-278. [38] B.P. Tripathi, N.C. Dubey, M. Stamm, Functional polyelectrolyte multilayer membranes for water purification applications, Journal of hazardous materials, 252–253 (2013) 401-412. [39] T. Carroll, N.A. Booker, J. Meier-Haack, Polyelectrolyte-grafted microfiltration membranes to control fouling by natural organic matter in drinking water, Journal of membrane science, 203 (2002) 3-13. [40] F. Diagne, R. Malaisamy, V. Boddie, R.D. Holbrook, B. Eribo, K.L. Jones, Polyelectrolyte and Silver Nanoparticle Modification of Microfiltration Membranes To Mitigate Organic and Bacterial Fouling, Environmental Science & Technology, 46 (2012) 4025-4033. [41] J. Choi, M.F. Rubner, Influence of the degree of ionization on weak polyelectrolyte multilayer assembly, Macromolecules, 38 (2005) 116-124. [42] W. Tong, C. Gao, H. Mohwald, Stable weak polyelectrolyte microcapsules with pHresponsive permeability, Macromolecules, 39 (2006) 335-340. 28
[43] Y.-H. Yang, L. Bolling, M. Haile, J.C. Grunlan, Improving oxygen barrier and reducing moisture sensitivity of weak polyelectrolyte multilayer thin films with crosslinking, RSC Advances, 2 (2012) 12355-12363. [44] P. Bieker, M. Schonhoff, Linear and exponential growth regimes of multilayers of weak polyelectrolytes in dependence on pH, Macromolecules, 43 (2010) 5052-5059.
Highlights • Membrane filtration coupled with electrocoagulation to treat wastewater. • Surface of NF 270 membrane modified by LbL of polyelectrolytes. • Modified membranes had higher flux and COD reduction than RO membranes. • Effect of effluent pH was studied. • Deposited films characterized based on surface charge and thickness.
29
PII: DOI: Reference:
S0376-7388(15)30175-7 http://dx.doi.org/10.1016/j.memsci.2015.09.011 MEMSCI13967
To appear in: Journal of Membrane Science Received date: 1 April 2015 Revised date: 11 August 2015 Accepted date: 6 September 2015 Cite this article as: Oishi Sanyal, Zhiguo Liu, Brooke M. Meharg, Wei Liao and Ilsoon Lee, Development of polyelectrolyte multilayer membranes to reduce the COD level of electrocoagulation treated High-strength wastewater, Journal of Membrane Science, http://dx.doi.org/10.1016/j.memsci.2015.09.011 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Development of polyelectrolyte multilayer membranes to reduce the COD level of electrocoagulation treated high-strength wastewater Oishi Sanyal1, Zhiguo Liu2, Brooke M. Meharg1, Wei Liao2 and Ilsoon Lee 1* 1
Department of Chemical Engineering and Materials Science, 2 Department of Biosystems and Agricultural Engineering Michigan State University, East Lansing, Michigan 48824, USA *To whom correspondence should be addressed: Email – [email protected]
Abstract This study focused on developing a membrane-based purification process, coupled with electrocoagulation (EC) as the pretreatment step, to reduce the COD level of an anaerobic digestion effluent. Commercial brackish water reverse osmosis (RO) membranes offer high COD removal but very low water fluxes. In an effort to address this issue, polyelectrolyte multilayer (PEM) membranes were fabricated by the surface modification of loose nanofiltration membranes using layer-by-layer assembly technique. The application of PEM membranes to treat wastewater effluents has not been explored in details. Two polyelectrolyte combinations were tried – the first one consisted of poly (diallyl dimethyl ammonium chloride) and poly (styrene sulfonate) while the second one consisted of poly (allylamine hydrochloride) and poly (acrylic acid). In comparison to commercial RO membranes, these membranes offered significantly higher fluxes, albeit with equivalent COD reduction. The effect of effluent properties like pH and composition, on the performance of these membranes has been discussed. The PEM films were characterized based on properties like thickness and surface charge, which directly affected the separation behavior of the membranes. For the first time, the combination of EC and PEM membranes has been tried out as a simple, energy-efficient two-step process for treating high-strength wastewater.
1
Keywords: High-strength wastewater; membrane filtration; COD reduction; layer-by-layer assembly; polyelectrolyte multilayer membranes
Introduction With rapid population increase and industrial growth, the fresh water resources are quickly being depleted. Improving the efficiency of water usage is critical to achieving global water sustainability. Recycle and reuse of municipal, industrial and agricultural wastewater are one of the key approaches to realize such efficiency improvement. Among different types of wastewater, high-strength wastewater with a significant amount of organic matter, solids, and nutrients requires more complicated processes to be treated and turned into clean water. In the present study, liquid effluent from an anaerobic digestion (AD) reactor was the target highstrength wastewater. Due to the chemical complexity of AD liquid effluent, Chemical Oxygen Demand (COD) was used as an indicator to evaluate the removal efficiency of different treatments [1, 2].
The combination of electrocoagulation (EC) and membrane technology has
attracted attention as an effective approach to treat high strength wastewater [3-6]. Compared to other physical and chemical treatments such as sedimentation, flocculation and activated carbon, EC technology presents a superior method to remove solids and pollutants from various wastewater streams. It has advantages of shorter retention time, better removal of smaller particles, no need of additional coagulation-inducing reagents, and smaller footprint [7]. A previous study demonstrated that a two-stage EC process showed high COD removal from AD effluent [8]. However, the EC treatment was still not enough to meet the drinking water standards. Thereby, in order to turn AD effluent into clean water, membrane filtration was
2
employed to purify the EC treated effluent. The overall wastewater treatment scheme is shown in Figure 1.
Figure 1. Flow diagram of the overall wastewater treatment process* *: Membrane filtration was the focus of this study.
Various membrane technologies such as microfiltration (MF), ultrafiltration (UF), and reverse osmosis (RO) have been implemented on EC treated wastewater to produce clean water [4, 9, 10].
Dense nanofiltration (NF) and RO membranes exhibit high solute removal
properties[11]. However, the permeability of the commercially available RO or dense NF membranes are extremely low, which leads to high operating costs. This study focused on the surface modification of NF 270 which is a “loose” NF membrane with high permeability. We 3
used the layer-by-layer assembly (LbL), which is an aqueous based thin film deposition technique, to modify the membrane surface. It was hypothesized that these membranes, on modification, would offer equal COD reduction as commercial RO membranes, but with higher solution permeability. The LbL technique was pioneered by Iler [12], and Decher et al. did a significant amount of research in this area to make LbL one of the most versatile thin film deposition techniques [13, 14]. It involves the deposition of polyelectrolytes which can interact via secondary molecular interactions such as ionic bonding, hydrogen bonding, hydrophobichydrophobic interactions etc. [15]. Polyelectrolyte multilayer (PEM) membranes, which are fabricated by the LbL assembly of polyelectrolytes on different commercial membrane surfaces, have been widely employed for ion rejection applications such as water softening or desalination [16-22]. Most of these studies were performed using a lab-based synthetic solution, which fails to emulate the complexities involved in real wastewater samples. Only a few studies applied the PEM-based modified membranes on real effluents such as industrial effluents and tanning effluents [23, 24]. In these studies, no detailed filtration performance and separation characteristics of the modified membranes have been reported. Besides, the combination of EC and PEM-based membrane filtration to treat high-strength agricultural wastewater also has not been studied. In order to elucidate the performance of PEM membranes two different sets of polyelectrolytes were selected and the resulting membranes were compared against a commercial RO membrane. The first set of polyelectrolytes comprised of the electrostatically crosslinked poly(diallyl dimethyl ammonium chloride) (PDAC) and poly(styrene sulfonate) (SPS), while the second set comprised of poly (allylamine hydrochloride) (PAH) and poly (acrylic acid) (PAA) that were covalently crosslinked by N-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC). COD
4
reduction and membrane permeability were used as the criteria to compare the performance of the PEM membranes with the commercial RO membranes.
2. Materials and Methods 2.1 Materials PDAC (MW 100,000-200,000, 20 wt% in water), SPS (MW 70,000), PAH (MW 900,000) and PAA (MW 15,000, 35 wt% in water) were purchased from Sigma Aldrich. 2-(NMorpholino) ethanesulfonic acid (MES) Buffer was also purchased from Sigma Aldrich. EDC was purchased from Fisher Scientific. Sodium chloride and potassium chloride crystals were procured from Avantor Performance Chemicals (Center Valley, PA). Three commercial membranes, NF 270, NF 90 and BW 30, from Dow Filmtec (Midland, MI), were used as the base membranes for this study. Among them, NF 270 and NF 90 are nanofiltration membranes, and BW 30 is a brackish water reverse osmosis membrane. All aqueous solutions were prepared using deionized (DI) water (>18.2 MΩ) supplied by a Barnstead Nanopure Diamond-UV purification unit equipped with a UV source and a final 0.2 µm filter. Unless specified all procedures were carried out at room temperature (25 °C). 2.2 The Dead End filtration set-up The HP 4750 stirred dead end cell (Sterlitech, Kent, WA) was connected to a nitrogen cylinder which acted as the pressure reservoir. The net volumetric capacity of the setup was around 300 ml. The unit was placed on a magnetic stirring plate. The filtrate was collected in a measuring cylinder. The flux was calculated by measuring the volume of water collected over a certain period of time and normalizing it with respect to the membrane area. The effective membrane area for this setup was 14.6 cm2.
5
2.3 Measurement of effluent properties The total solids (TS) measurement was done using the dry weight method. COD, total nitrogen (TN) and total phosphorus (TP) measurements were done using HACHTM standards methods [25]. Total carbon (TC) and inorganic carbon (IC) were measured using Shimadzu TOC-VCPN total organic carbon analyzer (Columbia, MD, US). Total organic carbon (TOC) was measured by subtracting the IC value from TC. 2.4 Experimental methods 2.4.1 Electrocoagulation (EC) treatment AD effluent was obtained from a 2500 m3 completely stirred tank reactor (CSTR) in the Anaerobic Digestion Research and Education Center (ADREC) at Michigan State University. A two-stage EC treatment was then carried out as described in a previous study [8]. Current level of 2A was applied to two sets of electrodes with anodic surface area of 62 cm2. The effective volume of the EC reactor was 0.5 L and retention time of 60 minutes was employed. The middle layer with relatively less turbidity was siphoned out as the feed for the second stage of EC, which applied same electricity conditions for another 40 minutes. Solutions were centrifuged at 236 g for 10 minutes before being used for the membrane filtration experiments. 2.4.2 Preparation of polyelectrolyte solutions and LbL assembly technique The bare NF 270 membranes were stored in DI water prior to the LbL deposition. The permeate sides of the membranes were covered with four alternate sheets of parafilm and aluminum foil prior to the LbL process. The required portion of the membrane was then cut out and used for the filtration experiments. This way, we were able to prevent the deposition of polyelectrolytes on both sides of the membrane to a large extent.
6
The LbL deposition process was carried out using a Carl Zeiss Slide Stainer which employs a robotic arm to move the sample to different solution baths. For all polyelectrolytes, the concentration was maintained at 10 mM. PDAC and SPS were prepared in 0.5M NaCl solutions and no pH adjustments were made to these solutions (the unadjusted pH values of the PDAC and SPS solutions were 6.6 and 6.0 respectively). For the PAH/PAA multilayer assembly, the pH of PAH and PAA were adjusted to 8.5 and 3.5, respectively, using 1M HCl/1M NaOH. The dipping time in each polyelectrolyte solution was set to 10 mins. After each polyelectrolyte dipping step, the substrates were rinsed with DI water for three
consecutive times (2 mins, 2
mins and 1 min). Following the deposition of one complete bilayer, the sample was sonicated for 2 mins in an ultrasonicator bath. The procedure was repeated till the desired number of bilayers was deposited. After LbL deposition, the membranes were soaked overnight in DI water. 2.4.3 Crosslinking of PAH and PAA with EDC The EDC solution was prepared in a 50 mM MES Buffer solution at a concentration of 50 mg/ml and pH of 5.5. Following the deposition of PAH/PAA multilayers on the membrane surface using the aforementioned LbL protocol, the modified membrane was dipped in the EDC solution using the slide stainer for 60 minutes with continuous agitation. The samples were then subject to three consecutive DI water rinsing steps, each of them being of 15 minutes duration, and then followed by 5 minutes of sonication. 2.4.4 Dead end filtration protocol The membranes were stored in DI water at least for 12 hours prior to usage. All membranes were first compacted by running pure DI water across them at a pressure of 10 bar. After two hours of compaction, the steady state value of the flux was noted down as the pure water flux of the membrane under consideration. Prior to using the wastewater, the solution was
7
pre-filtered using a 0.22 µm Millipore filter. This solution was then made to pass through the membranes at a transmembrane pressure (TMP) of 5 bar till the permeate volume reached 50 ml. The solution flux was calculated based on the time taken to filter 50 ml permeate across the given area of the membrane. The permeability and COD reduction values of various membranes which are presented in the later sections of this paper, were calculated based on the average of three replicates of each membrane. 2.5 Thin film characterization 2.5.1 Measurement of streaming potential The surface charge of the membranes were measured using the Brookhaven EKA Electro-kinetic analyzer (BI-EKA, Brookhaven Instrument Corp., Holtsville, NY) equipped with a rectangular cell and a clamp cell. A poly (methyl methacrylate) (PMMA) substrate was used as the reference. All streaming potential measurements were carried out using 1mM potassium chloride (KCl) as the electrolyte. The bypass lines and the cell were rinsed with KCl solution after being rinsed with DI water several times. At least three replicates were used for each of the membranes that were tested for streaming potential. In order to measure the membrane surface charge as a function of pH, the streaming potential tests were carried out with KCl solution at three different pH conditions (3, 7 and 9). The pH of these solutions was adjusted using 1M HCl/ 1M KOH. 2.5.2 Measurement of thickness The thickness of the PEM films was measured using a Dektak surface Profiler. Both types of PEM films were deposited on plain glass slides for the measurement. In order to reconfirm the values obtained from the profilometry tests, we also used the J.A Woollam M-44 Ellipsometer. The PEM films were deposited on gold coated glass slides (VWR International,
8
US) for the ellipsometry tests. Prior to the LbL deposition, the gold-coated glass slides as well as the plain glass slides were treated with O2 plasma for 20 minutes using a Harrick plasma cleaner (Harrick Scientific Corporation, Broading Ossining, NY) at 30 W RF power under 100 millitorr vacuum pressure. Immediately after the plasma treatment, the substrates were put in the slide stainer for the LbL process, as per the protocol described earlier (Section 2.4.1). The model for generic films was used for the ellipsometry measurements, assuming a refractive index of 1.5. The thickness was determined along several spots on the substrate and at least three replicates of each type of PEM film were used to get an average value.
. 3. Results and Discussions 3.1 EC treatment of diluted AD effluent The characteristics of AD liquid effluent and EC treated AD effluent (post vacuum filtration) have been listed in Table 1. After a two-stage EC treatment [8], a transparent solution was obtained and 97% of COD was removed, with the final COD level being 330 mg/L. Even though the EC treatment demonstrated excellent efficiency for solid and COD removal, the COD level in the effluent was still high, considering that the ultimate aim was to produce drinking water. Membrane-based filtration was therefore employed to reduce the COD of the EC treated AD effluent.
9
Table 1. Characteristics of the AD liquid effluent and filtered EC treated AD liquid effluent (post vacuum filtration) Parameters
AD liquid effluent*
EC-treated AD liquid effluent (post vacuum filtration) pH 7.5-8.0 9 TS (w/w %) 0.90 --1 COD (mg L ) 10017 330 TP (mg L-1) 340 1.18 TN (mg L-1) 1233 150 -1 TOC (mg L ) 2332 100 *: The AD liquid effluent for EC pretreatment was made by diluting the raw AD liquid digestate five times. 3.2 Performance of commercial membranes on the EC treated AD liquid effluent The three commercial membranes, NF 270, NF 90 and BW 30 were tested with the EC treated AD effluent and their performances were compared in terms of their permeability and COD reduction. The experimental data demonstrated that the NF 270 membrane had the highest pure water flux as well as solution flux among the three membranes, while BW 30 showed the highest COD removal (Figures 2 and 3). It has been reported that pH of the solution has significant influence on the separation behavior of the membranes [26]. The original pH of the EC treated AD effluent was around 9. The pH of the EC effluent was adjusted to 3 in order to evaluate the effect of pH on permeability and COD reduction. As shown in Figure 2, NF 270 membrane had higher COD reduction at acidic pH than alkaline pH, but the solution permeability was lower at acidic pH. (Figure 3a). In case of both NF 90 and BW 30 membranes, there was no significant difference in COD reduction under different pH conditions. The COD reduction of BW 30 membrane was higher than that of the NF 90 membrane. Both NF 90 and BW 30 had very similar solution permeability which was much lower than the solution permeability of NF 270 membrane. (Figure 3a). 10
The effect of pH on membranes, especially on NF 270 membrane, can be interpreted based on the membrane surface properties as well as the characteristics of the EC effluent. NF90 and BW30 are dense non-porous membranes as compared to NF 270 membrane, which is a “loose”, relatively more porous membrane. Even though all three membranes have positive charge at acidic conditions and negative charge at alkaline conditions, the responses of the NF 90 and BW 30 membranes to changes in effluent properties are different than that of NF 270 membrane. Natural organic matters (NOM) are present in almost all types of wastewater samples and the EC-treated AD effluent is no exception. TOC analysis confirmed the presence of a significant amount of organic matter in the EC effluent (100 mg/l). Among the various types of NOMs present in the solution, it has been reported that the humic compounds are largely responsible for the pH-dependent behavior of the effluent. [27-30]. As the humic compounds play critical roles as electron mediators in microbial electron transfer during anaerobic digestion [31], their concentration in the AD effluent was relatively high, and the EC treatment was not able to remove all of them. Consequently, a certain amount of them remained in the EC effluent. According to previous studies, humic compounds lose most of their charge under acidic conditions, resulting in their aggregation[29].
On the other hand, at alkaline pH these
compounds exhibit negative charge. For dense membranes like NF 90 and BW 30, the aggregation of humic compounds has little influence on their performance. Therefore no significant difference in performance (both COD removal and solution permeability) was observed for the NF 90 and BW 30 membranes, when tested under two different pH conditions. The BW 30 membranes showed the highest COD removal (92-95% ) with a COD level of
11
around 21 ppm in the final filtrate, but extremely low solution permeability of around 0.6 l/m2·hr·bar. For NF 270, the membrane surface properties as well as the EC effluent properties had significant influence on its performance. Under acidic conditions, the humic compounds were inclined to settle down on the membrane surface, resulting in lowering the permeability of NF 270 membrane at pH = 3 as compared to pH = 9 (Figure 3a). However, the deposition of humic compounds created an additional layer that prevented the passage of the remaining humic compounds and other fine particles in the effluent, which helped improve the COD reduction of the membrane (Figure 2). Under alkaline conditions, charge-based repulsion existed between the negatively charged humic compounds and the membrane surface. This repulsion was not enough to lower the COD level down to what was achieved under acidic conditions, but helped to maintain a higher solution permeability. Considering the fact, that a loose surface structure provides some flexibility for the LbL coating experiments, NF 270 was selected as the underlying substrate for fabricating the PEM membranes.
12
Figure 2. Comparison in COD reduction between the commercial membranes. (TMP = 5 bar)
(a)
13
(b) Figure 3. Permeability comparisons between the commercial membranes based on (a) solution tested under two different pH conditions and (b) pure water.
3.3 Performance of PEM-based membranes on EC-treated AD effluent 3.3.1 [PDAC (0.5 M NaCl)/SPS (0.5M NaCl)] 5.5 multilayer system Both PDAC and SPS are strong polyelectrolytes, which remain charged over a wide range of pH. 0.5 M NaCl was used as the supporting electrolyte which helps to increase the surface charge density [32]. The surface charge plays an important role in shaping the rejection behavior of these polyelectrolyte-based membranes. Under acidic conditions, as the humic compounds lose most of their charge, effect of the charge on the outermost layer is minimal. It was, however, important to understand how the surface charge affects the rejection under alkaline conditions, when the humic compounds are negatively charged. Two different systems were investigated to treat the EC effluent at pH 9: the first one comprising of 5 bilayers of
14
PDAC/SPS ending with SPS (negatively charged), and the second one consisting of 5.5 bilayers of PDAC/SPS ending with PDAC (positively charged). The experimental data demonstrated that the 5.5-bilayer system ending with PDAC had higher COD reduction (~ 84 %) than the system ending with SPS (~70 %) (Figure 4). Under alkaline conditions, the negatively charged humic compounds bind to the positively charged PDAC layer and form an additional barrier layer that helps enhance the rejection, while, the system ending with a SPS layer mainly uses charge-based repulsion to reject the negatively charged humic compounds. Therefore, under alkaline conditions, the barrier layer formation by humic compounds on a positively charged surface is more instrumental in reducing the COD level, than charge-based repulsion that occurs with a negatively charged outermost layer. Based on the performance results, the multilayer membrane ending with PDAC was chosen over the one ending with SPS, and used for the following experiments. The number of bilayers is also a tunable parameter for the LbL process. Following some preliminary experiments, 5.5 bilayers was found to be optimum in order to achieve both high flux as well as high rejection.
15
Figure 4. COD reduction of two PDAC/SPS based modified membranes. (TMP = 5 bar) (-) Bare NF270 or Multilayer assembly on NF270 ending with negatively charged SPS outer layer. (+) Multilayer assembly on NF270 ending with positively charged PDAC outer layer.
The performance of the PDAC/SPS–based membranes was further compared with the commercial membranes. These modified membranes showed around 84% COD reduction at pH = 9 and 90% at pH = 3 with corresponding COD values of 33 and 51.5 ppm in the filtered solution (Figure 5), which were significantly better than the original NF270. Permeability experiments demonstrated that the pure water permeability of the PDAC/SPS coated NF 270 membrane was much higher than NF 90 and BW 30 membrane (Figure 6b). The modified membrane also showed approximately 3 times higher solution permeability than BW 30/NF 90 at pH 3 and 5 times higher at pH 9 (Figure 6a). Similar to the underlying NF 270 membrane, the PDAC/SPS based membranes also had lower solution permeability at acidic pH as compared to alkaline pH (Figure 6a).
As mentioned before, the formation of a barrier layer of humic
compounds on the membrane surface was a key factor influencing the flux. However, the attachment mechanisms of the barrier layer under acidic and alkaline conditions are different. Under acidic conditions, the layer is formed due to the aggregation and settling of the humic compounds on the membrane surface. Under alkaline conditions, the negatively charged humic compounds adhere to the positively charged PDAC surface by electrostatic attraction. Given that the flux is lower under acidic condition, we presume that the barrier layer formed at pH 3 was thicker and denser than what was formed at pH 9. Owing to the thicker barrier layer, the rejection of the membrane was higher under acidic condition. The deposition of PDAC/SPS multilayers led to a considerable enhancement in the COD reduction compared to the bare NF 270 membrane. The percentage reduction was almost equal to that of BW 30 membrane under 16
acidic conditions but lower than BW 30 under alkaline conditions. Covalently cross-linked multilayer structure was therefore studied in order to further improve the membrane performance.
3.3.2 [PAH (pH 8.5)/PAA (pH 3.5)] 5.5 multilayer system Covalent cross-linking enables to create a tightly woven network, which is capable of blocking unwanted components of wastewater much more efficiently than ionic cross-linking. Therefore, a covalently cross-linked multilayer structure was coated on the surface of the NF 270 membrane with PAH and PAA as the surface modifiers. Both PAH and PAA are weak polyelectrolytes that exhibit pH-tunable behavior. When PAH is at pH 8.5 and PAA at pH 3.5, they form thick loopy films with inter-penetrated layers [33]. The amine groups of PAH and the carboxylic groups of PAA can be reacted to form covalent linkages between them. This can be simply done by heating the multilayers up to a high temperature (>180 °C) [34]. However, as the underlying NF 270 membrane cannot tolerate such high temperature, EDC induced chemical cross-linking was applied. [35]. A high EDC concentration of 50 mg/ml was used to ensure high cross-linking density [35]. In order to enable comparison with the PDAC/SPS –based membrane, the same number (5.5) of PAH/PAA bilayers were deposited. The performance of the PAH/PAA-based modified membrane has been presented in Figures 5 and 6. The COD reduction was between 91-95% under both acidic and alkaline conditions (Figure 5), and the final COD level in the permeate stream was almost equal to that of BW 30 membrane This could be attributed to a tight dense polymeric network that was formed by crosslinking PAH (pH 8.5) and PAA (pH 3.5), which was capable of completely
17
blocking the passage of unwanted components. The pure water permeability of the PAH/PAA membrane was the lowest among all the membranes (Figure 6b), while the solution permeability under both pH conditions was higher than BW 30 and NF 90 membranes (Figure 6a). The thick layers of PAH/PAA coating contributed to the low pure water permeability of the modified membrane. In spite of having a lower pure water permeability, the PAH/PAA coated NF 270 membrane showed 1.7 times higher solution permeability than the BW 30 membranes. We believe this was due to lower fouling propensity of the PAH/PAA coated membrane in comparison to the commercial RO membranes. This has been discussed in more details in a later section. In addition, because of the tight dense polymeric network and the thickness of the coated layers, the deposition of an additional layer of humic substances under acidic pH did not contribute much to the rejection. Therefore, there were no significant differences in COD reduction and solution permeability under acidic and alkaline conditions.
The PAH/PAA
multilayer membrane had the same COD reduction as the PDAC/SPS system at acidic pH, but higher reduction at alkaline pH The flux of this system was however, lower than the PDAC/SPS system. As the performance of the PAH/PAA coated NF 270 membrane was independent of the pH of the wastewater, it eliminates the need to acidify the solution prior to using it. This, in turn, helps to prevent the severe organic fouling that takes place under acidic conditions.
18
Figure 5. Comparison in COD reduction between the modified membranes and the commercial membranes at different pH conditions. (TMP = 5 bar)
19
(a)
(b) Figure 6. Permeability comparisons between the modified membranes and the commercial membranes based on (a) the solution at two different pH conditions and (b) pure water 20
The ratio of solution flux over pure water flux could be used as a criterion to evaluate the anti-fouling property of different membranes. A higher ratio indicates better fouling resistance offered by the membrane surface[36]. The PEM membranes (PDAC/SPS and PAH/PAA) showed better anti-fouling properties than their commercial counterparts (Table 2).The creation of anti-fouling surfaces by the deposition of polyelectrolytes has been reported earlier as well [37-40]. Among all the membranes, the PAH/PAA-based membranes had the best anti-fouling performance under both pH conditions.
Table 2. The ratio of the average solution flux to the average initial flux for commercial membranes and the PEM membranes* Membranes
Js (50)/Jw (0) at pH = 3
Js(50)/Jw(0) at pH = 9
NF 270
0.201
0.382
NF 270 modified with (PDAC/SPS)5.5 NF 270 modified with (PAH/PAA)5.5 NF 90
0.315
0.500
0.678
0.829
0.125
0.157
BW 30
0.209
0.193
Js(50) : The solution flux of the membrane measured by calculating the time taken to filter 50 ml of the permeate solution Jw(0) : The pure water flux or the initial flux of the membrane *: Both the fluxes were normalized with respect to the pressure applied during the measurements
21
With the combination of EC and membrane filtration, the color of the AD effluent was completely removed (Figure 7). This is important from the point of view of producing potable water. The EC process greatly removed the majority of the COD and TS in the AD effluent. The following membrane filtration step, irrespective of the membrane used, further decolored and cleaned the EC treated AD effluent. A relatively less energy-demanding method for treating high-strength wastewater was therefore established.
Figure 7. Color change of the solution following different treatment stages Left to right: AD effluent, EC effluent and membrane filtrate.
3.4 Effect of the coating properties on the membrane performance 3.4.1 Thickness of the PEM coatings The thickness of the coatings plays an important role in determining the permeability of the modified membrane. The strong polyelectrolytes PDAC and SPS are fully charged irrespective of pH conditions. The strong ionic linkages between the positively charged PDAC backbones and negatively charged SPS lead to the formation of very thin flat coatings. The 22
addition of NaCl increases the surface charge density and also increases the thickness of the films [20]. PAH and PAA exhibit pH dependent ionization behavior. PAH is around 50% ionized at a pH of 8.5, and PAA is only about 10% ionized at pH of 3.5 [41]. Since they are only partially charged, they arrange themselves in a coiled conformation,
with a high level of
interlayer diffusion [33], which leads to the formation of thick loopy films.. Crosslinking reduced the swelling tendency of the polyelectrolytes and in turn also led to a decrease in the film thickness.[42, 43]. The thickness values of the two types of polyelectrolyte systems have been shown in Table 3. It is apparent that the thickness of the coatings directly affects the permeability. The thicker PAH/PAA coatings had lower initial permeability than the PDAC/SPSbased coatings.
Table 3. The thickness values of the two types of polyelectrolyte coatings on NF 270 membrane Type of coating
Thickness (nm)
[PDAC (0.5M NaCl) / SPS(0.5M NaCl)] 5.5
28.70 ± 4.66
[PAH (pH 8.5) / PAA (pH 3.5)]5.5
134.15 ± 9.87
3.4.2 Surface charge of the PEM coatings Surface streaming potential of membranes plays an important role in determining the membrane rejection properties. The streaming potentials of all three commercial membranes as well as two modified membranes under three different pH conditions (acidic, neutral and alkaline) are listed in Table 4.
23
Table 4. The streaming potential data for the commercial membranes and the modified membranes as a function of pH* Surface streaming potential (mV) @ pH 3
Surface streaming potential (mV) @ pH 7
Surface -streaming potential (mV) @ pH 9
NF 270
5.54 ± 0.13
- 4.17 ± 1.89
-21.35 ± 0.36
NF 270 modified with (PDAC/SPS)5.5 NF 270 modified with (PAH/PAA)5.5 NF 90
20.38 ± 1.66
29.24 ± 2.45
20.07 ± 5.86
20.50 ± 4.47
12.83 ± 2.30
10.26 ± 3.36
20.91 ± 0.99
- 15.74 ± 0.09
-15.33 ± 0.23
BW 30
12.70 ± 3.80
- 4.70 ± 1.45
-3.66 ± 0.45
Membrane
*: Data represent average of three replicates with standard deviation.
The PEM membranes remain positively charged throughout the entire pH range of 3-9, unlike the commercial membranes. For the PDAC/SPS system, the outermost PDAC layer remained almost fully ionized within a wide range of pH and therefore imparted positive charge to the membrane surface. This leads to a small variation in its performance under different pH conditions (Table 4). For the PAH/PAA system, the effect of pH is much more complicated since both polyelectrolytes have pH dependent ionization. When exposed to a pH of 3, the outermost PAH layer was fully ionized and therefore had the highest positive charge. With the decrease in percentage ionization, the surface streaming zeta potential also decreased as the pH was increased (Table 4). However, many other complicated phenomena are associated with this polyelectrolyte system, like the inter-diffusion of polyelectrolytes [44] and shifting of the pKa value of the polyelectrolytes [41] when incorporated within the multilayers.. Besides, the effect of cross-linking on the ionization of the weak polyelectrolytes not been studied till date. These
24
factors can individually or collectively influence the surface streaming potentials under varying pH conditions.
4. Conclusion With this study it can be concluded that integration of electrocoagulation and PEM-based membrane separation provides an efficient way to reclaim a high-strength wastewater. PEMbased membranes demonstrated excellent performance in treating an actual wastewater effluent. A comprehensive study of several key concepts involving the membrane surface properties as well as the effluent properties was performed to deal with the complexities involved in wastewater treatment. The pH of the wastewater solution played a pivotal role in determining the performance of the membranes mainly due to the NOMs present in the EC effluent. Both PDAC/SPS and PAH/PAA based membranes had higher solution permeability than the commercial reverse osmosis membrane (BW 30). The PDAC/SPS based membrane showed equivalent COD reduction as BW 30 under acidic conditions, while the PAH/PAA based membranes showed equal reduction as BW 30, irrespective of the pH conditions used. Both the polyelectrolyte-based membranes had much lower fouling propensity than the commercial membranes, which expands the scope of applicability of PEM membranes beyond some simple ion-rejection studies. List of Figure Captions 1. Flow diagram of the overall wastewater treatment process* *: Membrane filtration was the focus of this study 2. Comparison in COD reduction between commercial membranes. (TMP = 5 bar) 3. Permeability comparisons between the commercial membranes based on (a) solution tested under two different pH conditions and (b) pure water
25
4. COD reduction of two PDAC/SPS based modified membranes (TMP = 5 bar) (-) Bare NF 270 or Multilayer assembly on NF 270 ending with negatively charged SPS outer layer (+)Multilayer assembly on NF 270 ending with positively charged PDAC outer layer 5. Comparison in COD reduction between the modified membranes and the commercial membranes at different pH conditions. (TMP = 5 bar) 6. Permeability comparisons between the modified membranes and the commercial membranes based on (a) the solution at two different pH conditions and (b) pure water. 7. Color change of the solution following different treatment stages Left to right: AD effluent, EC effluent and membrane filtrate
Acknowledgment The authors sincerely thank Prof. Volodymyr Tarabara from the Department of Civil and Environmental Engineering at MSU for allowing the usage of EKA Electrokinetic analyzer in his lab. The authors would also like to acknowledge the DOD Strategic Environmental Research and Development Program (DOD SERDP W912HQ-12-C-0020) for funding the entire research work.
References [1] H.A. Moreno-Casillas, D.L. Cocke, J.A.G. Gomes, P. Morkovsky, J.R. Parga, E. Peterson, Electrocoagulation mechanism for COD removal, Separation and Purification Technology, 56 (2007) 204-211. [2] K. Yetilmezsoy, F. Ilhan, Z. Sapci-Zengin, S. Sakar, M.T. Gonullu, Decolorization and COD reduction of UASB pretreated poultry manure wastewater by electrocoagulation process: A posttreatment study, Journal of hazardous materials, 162 (2009) 120-132. [3] M. Ben-Sasson, Y.M. Lin, A. Adin, Electrocoagulation-membrane filtration hybrid system for colloidal fouling mitigation of secondary-effluent, Separation and Purification Technology, 82 (2011) 63-70. [4] M. Ben-Sasson, Y. Zidon, R. Calvo, A. Adin, Enhanced removal of natural organic matter by hybrid process of electrocoagulation and dead-end microfiltration, Chemical Engineering Journal, 232 (2013) 338-345. [5] V. Mavrov, S. Stamenov, E. Todorova, H. Chmiel, T. Erwe, New hybrid electrocoagulation membrane process for removing selenium from industrial wastewater, Desalination, 201 (2006) 290-296. 26
[6] V. Vinduja, N. Balasubramanian, Electrocoagulation-integrated hybrid membrane processes for the treatment of tannery wastewater, Environmental Science and Pollution Research, 20 (2013) 7441-7449. [7] M.Y.A. Mollah, R. Schennach, J.R. Parga, D.L. Cocke, Electrocoagulation (EC)- science and applications, Journal of hazardous materials, 84 (2001) 29-41. [8] Z. Liu, D. Stromberg, X. Liu, W. Liao, Y. Liu, A new multiple-stage electrocoagulation process on anaerobic digestion effluent to simultaneously reclaim water and clean up biogas, Journal of hazardous materials, 285 (2015) 483-490. [9] X. Chen, H. Deng, Removal of humic acids from water by hybrid titanium-based electrocoagulation with ultrafiltration membrane processes, Desalination, 300 (2012) 51-57. [10] W. Den, C.-J. Wang, Removal of silica from brackish water by electrocoagulation pretreatment to prevent fouling of reverse osmosis membranes, Separation and Purification Technology, 59 (2008) 318-325. [11] C. Bellona, J.E. Drewes, P. Xu, G. Amy, Factors affecting the rejection of organic solutes during NF/RO treatment—a literature review, Water Research, 38 (2004) 2795-2809. [12] R.K. Iler, Multilayers of colloidal particles, Journal of Colloid and Interface Science, 21 (1966) 569-594. [13] G. Decher, Fuzzy nanoassemblies: Toward layered polymeric multicomposites, Science, 277 (1997) 1232-1237. [14] G. Decher, J.D. Hong, J. Schmitt, Buildup of Ultrathin Multilayer Films by a Self-Assembly Process .3. Consecutively Alternating Adsorption of Anionic and Cationic Polyelectrolytes on Charged Surfaces, Thin Solid Films, 210 (1992) 831-835. [15] I. Lee, Molecular Self-Assembly: Smart Design of Surface and Interface via Secondary Molecular Interactions, Langmuir, 29 (2013) 2476-2489. [16] S.U. Hong, O.Y. Lu, M.L. Bruening, Recovery of phosphate using multilayer polyelectrolyte nanofiltration membranes, Journal of membrane science, 327 (2009) 2-5. [17] S.U. Hong, R. Malaisamy, M.L. Bruening, Optimization of flux and selectivity in Cl-/SO42separations with multilayer polyelectrolyte membranes, Journal of membrane science, 283 (2006) 366-372. [18] S.U. Hong, R. Malaisamy, M.L. Bruening, Separation of fluoride from other monovalent anions using multilayer polyelectrolyte nanofiltration membranes, Langmuir, 23 (2007) 17161722. [19] L. Krasemann, B. Tieke, Selective ion transport across self-assembled alternating multilayers of cationic and anionic polyelectrolytes, Langmuir, 16 (2000) 287-290. [20] L. Ouyang, R. Malaisamy, M.L. Bruening, Multilayer polyelectrolyte films as nanofiltration membranes for separating monovalent and divalent cations, Journal of membrane science, 310 (2008) 76-84. [21] O. Sanyal, I. Lee, Recent progress in the application of Layer-by-layer assembly to the preparation of nanostructured ion-rejecting water purification membranes, Journal of Nanoscience and Nanotechnology, 14 (2014) 2178-2189. [22] O. Sanyal, A.N. Sommerfeld, I. Lee, Design of ultrathin nanostructured polyelectrolytebased membranes with high perchlorate rejection and high permeability, Separation and Purification Technology, 145 (2015) 113-119. [23] U.K. Aravind, B. George, M.S. Baburaj, S. Thomas, A.P. Thomas, C.T. Aravindakumar, Treatment of industrial effluents using polyelectrolyte membranes, Desalination, 252 (2010) 2732. 27
[24] M.S. Baburaj, C.T. Aravindakumar, S. Sreedhanya, A.P. Thomas, U.K. Aravind, Treatment of model textile effluents with PAA/CHI and PAA/PEI composite membranes, Desalination, 288 (2012) 72-79. [25] A.M. Jirka, M.J. Carter, Analytical Chemistry, 47 (1975) 1397. [26] G. Capar, L. Yilmaz, U. Yetis, Reclamation of acid dye bath wastewater: effect of pH on nanofiltration performance, Journal of membrane science, 281 (2006) 560-569. [27] S. Hong, M. Elimelech, Chemical and physical aspects of natural organic matter (NOM) fouling of nanofiltration membranes, Journal of membrane science, 132 (1997) 159-181. [28] C.Y. Tang, Y.-N. Kwon, J.O. Leckie, Fouling of reverse osmosis and nanofiltration membranes by humic acids-Effects of solution composition and hydrodynamic conditions, Journal of membrane science, 290 (2007) 86-94. [29] W. Yuan, A.L. Zydney, Effects of solution environment on humic acid fouling during microfiltration, Desalination, 122 (1999) 63-76. [30] W. Yuan, A.L. Zydney, Humic acid fouling during microfiltration, Journal of membrane science, 157 (1999) 1-12. [31] L. Shao, T. Wang, T. Li, F. Lu, P. He, Comparison of sludge digestion under aerobic and anaerobic conditions with a focus on the degradation of proteins at mesophilic temperature, Bioresource technology, 140 (2013) 131-137. [32] B.W. Stanton, J.J. Harris, M.D. Miller, M.L. Bruening, Ultrathin, multilayered polyelectrolyte films as nanofiltration membranes, Langmuir, 19 (2003) 7038-7042. [33] S.S. Shiratori, M.F. Rubner, pH-dependent thickness behavior of sequentially adsorbed layers of weak polyelectrolytes, Macromolecules, 33 (2000) 4213-4219. [34] J. Park, J. Park, S.H. Kim, J. Cho, J. Bang, Desalination membranes from pH-controlled and thermally-crosslinked layer-by-layer assembled multilayers, Journal of Materials Chemistry, 20 (2010) 2085-2091. [35] P. Schuetz, F. Caruso, Copper-Assisted Weak Polyelectrolyte Multilayer Formation on Microspheres and Subsequent Film Crosslinking, Advanced Functional Materials, 13 (2003) 929-937. [36] J. Pieracci, J.V. Crivello, G. Belfort, Photochemical modification of 10kDa polyethersulfone ultrafiltration membranes for reduction of biofouling, Journal of membrane science, 156 (1999) 223-240. [37] W. Shan, P. Bacchin, P. Aimar, M.L. Bruening, V.V. Tarabara, Polyelectrolyte multilayer films as backflushable nanofiltration membranes with tunable hydrophilicity and surface charge, Journal of membrane science, 349 (2010) 268-278. [38] B.P. Tripathi, N.C. Dubey, M. Stamm, Functional polyelectrolyte multilayer membranes for water purification applications, Journal of hazardous materials, 252–253 (2013) 401-412. [39] T. Carroll, N.A. Booker, J. Meier-Haack, Polyelectrolyte-grafted microfiltration membranes to control fouling by natural organic matter in drinking water, Journal of membrane science, 203 (2002) 3-13. [40] F. Diagne, R. Malaisamy, V. Boddie, R.D. Holbrook, B. Eribo, K.L. Jones, Polyelectrolyte and Silver Nanoparticle Modification of Microfiltration Membranes To Mitigate Organic and Bacterial Fouling, Environmental Science & Technology, 46 (2012) 4025-4033. [41] J. Choi, M.F. Rubner, Influence of the degree of ionization on weak polyelectrolyte multilayer assembly, Macromolecules, 38 (2005) 116-124. [42] W. Tong, C. Gao, H. Mohwald, Stable weak polyelectrolyte microcapsules with pHresponsive permeability, Macromolecules, 39 (2006) 335-340. 28
[43] Y.-H. Yang, L. Bolling, M. Haile, J.C. Grunlan, Improving oxygen barrier and reducing moisture sensitivity of weak polyelectrolyte multilayer thin films with crosslinking, RSC Advances, 2 (2012) 12355-12363. [44] P. Bieker, M. Schonhoff, Linear and exponential growth regimes of multilayers of weak polyelectrolytes in dependence on pH, Macromolecules, 43 (2010) 5052-5059.
Highlights • Membrane filtration coupled with electrocoagulation to treat wastewater. • Surface of NF 270 membrane modified by LbL of polyelectrolytes. • Modified membranes had higher flux and COD reduction than RO membranes. • Effect of effluent pH was studied. • Deposited films characterized based on surface charge and thickness.
29